full_analysis

R Markdown

Here is a complete Markdown document combining work from ER, LN and DG’s combined analysis on Rare’s Fish Forever survey data.

To start, we’ll read in all relevant CSVs

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   .default = col_double(),
##   submissionid = col_character(),
##   country = col_character(),
##   level1_name = col_character(),
##   level2_name = col_character(),
##   level3_name = col_logical(),
##   level4_name = col_character(),
##   ma_name = col_character(),
##   ma_status = col_character(),
##   reserve_status = col_character(),
##   username = col_character(),
##   `1_interviewer` = col_character(),
##   `2_affiliation` = col_character(),
##   `3_community` = col_character(),
##   `5_hh_status` = col_character(),
##   `6_gender` = col_character(),
##   `8_religion` = col_character(),
##   `8_religion_other` = col_character(),
##   `9_region_name` = col_character(),
##   `11k_other_source` = col_character(),
##   `14_responsibilities_other` = col_character()
##   # ... with 56 more columns
## )
## ℹ Use `spec()` for the full column specifications.

## Warning: 24212 parsing failures.
##  row                  col           expected        actual                                                                                                    file
## 1023 level3_name          1/0/T/F/TRUE/FALSE Maginti       '/Users/dylan/Desktop/Summer W: Rare and Tasha/FishForever_Research/new_portal_data/hh_surveys_all.csv'
## 1024 level3_name          1/0/T/F/TRUE/FALSE Kabaena Barat '/Users/dylan/Desktop/Summer W: Rare and Tasha/FishForever_Research/new_portal_data/hh_surveys_all.csv'
## 1025 level3_name          1/0/T/F/TRUE/FALSE Tiworo Utara  '/Users/dylan/Desktop/Summer W: Rare and Tasha/FishForever_Research/new_portal_data/hh_surveys_all.csv'
## 1025 35f_ma_gear_mosquito 1/0/T/F/TRUE/FALSE Unknown       '/Users/dylan/Desktop/Summer W: Rare and Tasha/FishForever_Research/new_portal_data/hh_surveys_all.csv'
## 1025 35g_ma_gear_poison   1/0/T/F/TRUE/FALSE Not permitted '/Users/dylan/Desktop/Summer W: Rare and Tasha/FishForever_Research/new_portal_data/hh_surveys_all.csv'
## .... .................... .................. ............. .......................................................................................................
## See problems(...) for more details.

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   submissionid = col_character(),
##   ma_id = col_double(),
##   `44_meeting_attendance` = col_character()
## )

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   submissionid = col_character(),
##   ma_id = col_double(),
##   `45_leadership_position` = col_character()
## )

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   submissionid = col_character(),
##   ma_id = col_double(),
##   `48_enforcement_participation` = col_character()
## )

Next, we go through questions 44, 45, and 48 to remove any duplicates

We do this because, when reading in the original CSVs, we saw answers like, “Yes, Yes, Yes,” “No, No, No, No, No” and “Yes male, yes Female, Unsure.”

These questions ask respondents about their household, so it’s possible that they answered individually for each member of their household. However, the goal of this analysis is just to see if any single member of a household participated the management meetings, or enforcement.

The chunks below use recode_factor() to change cells within Qs 44, 45, and 48 to each have a single answer.

## `summarise()` has grouped output by 'submissionid'. You can override using the `.groups` argument.

The following chunk combines duplicated from Question 45, and renames the observations that have been repeated. End result: 1 observation for each submission ID

Note: After Larissa did all the cleaning below, there was an issue witht the recode() function. Dylan replaced each of these with recode_factor() and all factor levels are fine.

## `summarise()` has grouped output by 'submissionid'. You can override using the `.groups` argument.

Below, the same happens for question 48

## `summarise()` has grouped output by 'submissionid'. You can override using the `.groups` argument.

Again, recode() has been switched to recode_factor() for question 48

After recoding/cleaning, these different df were all combined into hhs_complete_all using right_join

## Warning: Missing column names filled in: 'X1' [1]

## 
## ── Column specification ────────────────────────────────────────────────────────
## cols(
##   .default = col_double(),
##   submissionid = col_character(),
##   `44_meeting_attendance` = col_character(),
##   `45_leadership_position` = col_character(),
##   `48_enforcement_participation` = col_character(),
##   country = col_character(),
##   level1_name = col_character(),
##   level2_name = col_character(),
##   level3_name = col_logical(),
##   level4_name = col_character(),
##   ma_name = col_character(),
##   ma_status = col_character(),
##   reserve_status = col_character(),
##   username = col_character(),
##   `1_interviewer` = col_character(),
##   `2_affiliation` = col_character(),
##   `3_community` = col_character(),
##   `5_hh_status` = col_character(),
##   `6_gender` = col_character(),
##   `8_religion` = col_character(),
##   `8_religion_other` = col_character()
##   # ... with 59 more columns
## )
## ℹ Use `spec()` for the full column specifications.

Here is a quick summary table of the total number of responses, which matches the original number of surveys completed 17712

Filtering Cleaned Data

Alright!

This marks the end of the “cleaning” data section. The final two dataframes (hhs_all_complete and summary) have each survey response as 1 row, with no duplicates, and questions that were answered multiple times (44, 45, and 48) now have 1 answer of Yes, No, or Unsure.

Filtering out NAs:

Goal: A robust df with only observations (survey respondents) who have answered every engagement and agreement question.

This will be called hhs_complete_filtered

First, let me get counts of the NAs that we are changing to neutral for Q 44 and 45. We think these two will filter out a large amount of observations:

We see 10,000 NA observations. This is ok, since many of those observations contain NAs for other varables of interest. I.e: they’d be weeded out eventually.

Here is the creation of our filtered df, where each observation responds to every agreement and engagement question. These questions exist in rows 24:33, therefore drop_na is used for those columns.

Update on Agreement From 9/9/2021: In analysis with Dr. Quintana, after looking at variable independence and reframing of what Agreement includes, we are removing the following three questions from the Index: - Question 10 (x10_mpa_important) due to wording that doesn’t focus on MPAs - Question 47 (x47_represent_contributions) due to its near exact math to question 46. This would give these two similar questions as much weight as the other agreement components, which we don’t want. - Question 62 (x62_reserve_compliance) because this question outright asks if fishers follow the rules. This does not have to do with their Agreement - Question 64 (x64_wrong_fishing_reserve) was not asked in Brazil, and we want to include these 756 observations, as the country is a lage focus of the Fish Forever process.

Now columns [24:33] within hhs_complete_filtered are all of the Engagement and Agreement questions.

• Dropping NAs: We drop NAs for all Engagement and Agreement Questions.

• For Q 46-47 we’ve excluded NAs since they mostly represent “non-fisher households” The difference in n goes from 5640 to ~3300. We want to see how fishers comply though. So the smaller N will show us that.

Quantitiative values of survey Answers

Let me list out which A and E questions have binary responses, for those that have “neither agree nor disagree,” then I’ll move from -1 to 1, with a neutral option 0.

Agreement

I’ll start converting the straightforward questions to 1s and 0s. Because there was so much much dang coding in the global_filtered, I’ll make a new df for the numbered version. We’ll call it hhs_filtered_numeric**

Engagement

Quantification list:

• 44: Yes/No 1/-1 (We recoded not sure to 0, see Methods)

• 45: Yes/No 1/-1 (We recoded ’Not sure to 0, see Methods)

• 48: Yes/No/ Not Sure 1/0/-1

• 53: Very Often - Never [As a back up, this would be flagged for this being its own index of engagement]

• 61 (f): Likert Scale (Strongly) Disagree-Agree (1-5)

• 61 (h): Likert Scale (Strongly) Agree-Disagree

Agreement

• 10: Yes/No/Neutral 1,0,-1
• 43: Yes/No: 1, 0, -1
• 46: Agree/Neither/Disagree: 1, 0. -1
• 47: Agree/Neither-No Management - na/Disagree: 1, 0. -1
• 52: Yes/Unsure-No management-na/No: 1, 0,-1
• 61(a): Likert Scale (Strongly) Agree-Disagree 0-5
• 61(e): Likert Scale (Strongly) Agree-Disagree 0-5
• 64: How wrong is fishing in the reserve (Not bad at all, Very bad): 1-5

The chunk below uses mutate() and recode to complete the numeration above.

Notes: - For the Likert scale questions that Rare coded from 0-5, I removed the 0s, since they were equivalent to NAs.

All “No management” or “No regulations” ar gone. and all 0s have been removed from questions that rare coded from 0-5. Only 1 question made a significant change, x61f_rights_distribution_fair

Now, I believe it’s time to scale the data!

Let’s just make a quick table with the counts of all the survey responses that we have for analysis.

Engagement and Agreement Summary

Managed Access Area	Number of surveys with all ‘Agreement’ and ‘Engagement’ questions answered
SDN_SantaMonica	256
Antique_Libertad	238
SDN_SanBenito	230
SDN_DelCarmen	227
Antique_Culasi	222
SDN_GL	209
Antique_Sebaste	208
SDN_SanIsidro	195
SDN_Pilar	191
RESEX Marinha Mocapajuba	184
SDN_Burgos	167
SDN_Socorro	160
RESEX Soure	149
SDN_Dapa	148
RESEX Mãe Grande de Curuça	143
Santa Fé	129
Antique_Pandan	106
RESEX São João da Ponta	103
RESEX Gurupi-Piriá	94
Baixo Pinda	92
Trujillo	92
Teluk Kolono	88
Zavora	84
Pomene	83
Fequete	82
Memba	66
Serissa	63
Guanaja	59
Machangulo	54
NA	47
Tsondzo	37
Mucocuene	34
Wangi-Wangi Selatan	34
Kitti	28
Tomia	24
Farol	20
Ngiwal	19
Melekeok	18
Nhagondzo	18
Vuca	18
Petane	15
Cebu_SanFrancisco	7
Simuco	4
Iriona	2
Puerto Cortés	1
RESEX Caeté-Taperaçu	1
Total	4449

Counts by Reserve is great, but Larissa would like the same table by Country

Engagement and Agreement Country Summary

Country	Number of surveys with all ‘Agreement’ and ‘Engagement’ questions answered
PHL	2564
BRA	674
MOZ	670
HND	330
IDN	146
PLW	37
FSM	28
Total	4449

Scaling

I’ve double-checked the statistics on each of the above variables! Each new column has a mean of nearly 0 (3x10^-16, etc.)

Since we changed recode() to recode_factor(), these stats functions would not run. By changing the values to as.numeric, all is well.

Graphing Indices

Let’s start with mean() and sum()

Along with scaling the data onto a Z-Axis, I’m going to simply sum across the agreement and engagement columns too, to see what we get! - eng_mean is the mean index score of engagement, with 1 score for each fisher respondent - agree_mean is the mean score of agreement for each fisher respondent - I add two more variables: “raw_eng_sum” and “raw_agree_sum.” Without scaling, I’ve summed agreement and engagement across rows, for exploratory visualization

These sum variables are exploratory, with the mean score for each observation is the index that we will use.

Here’s a Table of Index scores

Engagement and Agreement Index Scores

country	Relative Engagement Score	Relative Agreement Score
PHL	0.9644028	-0.7650206
PHL	-0.4500038	-0.4231813
PHL	0.1720092	0.0118288
BRA	-0.6392887	0.0993151
HND	-0.8285736	-2.1032934
PHL	0.0517048	0.3221292

With that, we’ll make a preliminary graph:

## Warning: Removed 49 rows containing missing values (geom_point).

## Warning: Removed 25 rows containing missing values (geom_point).

Here is a graph of Agreement and Engagement measures, colored by responses of “Do you fish inside the reserve?”

Question 63 asks if fishers break the rules, and fish in the reserve. This would be a classic measure of compliance (although self-reporting likely leads to a lower N of admitted non-compliers)

## Warning: Removed 2 rows containing missing values (geom_point).

I want to see how many people said “yes” here:

138 fishers admitted that they fish in the reserve, while 2800 said they do not. The question we’re left with is: What is different about folks who admit to breaking rules? These “Yes” rule-breakers are spread relatively evenly across all countries. Maybe they don’t think there are sanctions, their community doesn’t care about it

Brazil analysis We now include 756 Brazilian observations

• There are 1592 observations from Brazil
• When we filter out NAs for engagement questions, there are 1018
• When we filter NAs for question 10 and 43, we have 1008 (no real difference)
• Filtering 46, 47, 52, 43, 61a, 61f and we drop to 811 (still not bad)
• but there is only ONE! observation from brazil that answered questions 62 and 64 (I wish we’d asked Courtney about this)

Can we (responsibly) mark these folks as neutral? NOPE mena of 64: 3.9/5 (Most fishers list fishing in the reserve as ‘bad’ or ‘very bad’)

• Is it Okay to just model z-scores? We want to see actual levels of compliance, but this is the first analysis of its type. We do not have objective “standards” where a fisher’s response would change from reluctant to committed.

We’ll run two different LN/Logit Models: one for agreement, one for engagement

• Linear regression will give us coefficients for categorical and ordinal questions, but I’ll clean the bundles of questions that we want to analyze.

[https://stats.idre.ucla.edu/r/dae/multinomial-logistic-regression/]

Statistical Summaries of Agreement and Engagement

Making some tables and graphs for Claudia + Rare:

Summary Statistics

Variable	N	Mean	Std. Dev.	Min	Pctl. 25	Pctl. 75	Max
x44_meeting_attendance	4449	0.053	0.969	-1	-1	1	1
x45_leadership_position	4449	-0.495	0.824	-1	-1	0	1
x48_enforcement_participation	4449	-0.533	0.774	-1	-1	0	1
x53_encourage_regulations	4449	-0.022	1.057	-2	-1	1	2
x61g_fishing_change_behavior	4449	3.975	0.822	1	4	4	5

Summary Statistics

Variable	N	Mean	Std. Dev.	Min	Pctl. 25	Pctl. 75	Max
x43_ma_benefits	4449	0.57	0.498	-1	0	1	1
x46_represent_interests	4449	0.649	0.594	-1	0	1	1
x52_ma_benefit_5yrs	4449	0.502	0.582	-1	0	1	1
x61a_current_regulations	4449	3.875	0.898	1	4	4	5
x61f_rights_distribution_fair	4449	3.77	0.92	1	3	4	5

Great!

First, hhs_scaled is the df that will be used for all analysis.

I’ve run unique() for all the E & A questions, ensuring that we only have responses that are measurable.

Following this with a check of the “no management” answers, and how they were coded.

With the statistics, we see the new mean, sd, etc. for all the agreement and engagement question.

Playing around with the Logit Model

Fully Exploratory

Do not sue the following as results

## 
## Call:
## lm(formula = eng_mean ~ x50_ma_punishment + x40_reserve_boundaries_aware, 
##     data = scaled_sum)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.66591 -0.44466 -0.09244  0.38738  1.87105 
## 
## Coefficients:
##                                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                    -0.281473   0.028430  -9.901  < 2e-16 ***
## x50_ma_punishmentna            -0.170936   0.130346  -1.311  0.18979    
## x50_ma_punishmentNo management  0.362935   0.125148   2.900  0.00375 ** 
## x50_ma_punishmentNo punishment -0.003160   0.040209  -0.079  0.93737    
## x50_ma_punishmentSevere         0.280443   0.032277   8.689  < 2e-16 ***
## x50_ma_punishmentStrong         0.112347   0.021542   5.215 1.93e-07 ***
## x50_ma_punishmentWeak           0.113195   0.039381   2.874  0.00407 ** 
## x40_reserve_boundaries_aware    0.027613   0.002843   9.711  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.593 on 4080 degrees of freedom
##   (361 observations deleted due to missingness)
## Multiple R-squared:  0.04593,    Adjusted R-squared:  0.0443 
## F-statistic: 28.06 on 7 and 4080 DF,  p-value: < 2.2e-16

## 
## Call:
## lm(formula = eng_mean ~ x19_current_fish_catch + x22_catch_5yrs + 
##     x23_job_secure, data = scaled_sum)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.69350 -0.41150 -0.09613  0.37513  1.80737 
## 
## Coefficients:
##                                          Estimate Std. Error t value Pr(>|t|)
## (Intercept)                             -0.001244   0.025151  -0.049   0.9606
## x19_current_fish_catchDeclined slightly -0.057742   0.027124  -2.129   0.0333
## x19_current_fish_catchImproved heavily   0.181116   0.118112   1.533   0.1252
## x19_current_fish_catchImproved slightly  0.062014   0.040817   1.519   0.1288
## x19_current_fish_catchStayed the same   -0.196630   0.030854  -6.373 2.05e-10
## x22_catch_5yrsDeclines slightly         -0.037434   0.030021  -1.247   0.2125
## x22_catch_5yrsImproves heavily           0.235881   0.095265   2.476   0.0133
## x22_catch_5yrsImproves slightly          0.155299   0.035652   4.356 1.36e-05
## x22_catch_5yrsStays the same            -0.023083   0.031156  -0.741   0.4588
## x23_job_secure                           0.111263   0.019129   5.816 6.46e-09
##                                            
## (Intercept)                                
## x19_current_fish_catchDeclined slightly *  
## x19_current_fish_catchImproved heavily     
## x19_current_fish_catchImproved slightly    
## x19_current_fish_catchStayed the same   ***
## x22_catch_5yrsDeclines slightly            
## x22_catch_5yrsImproves heavily          *  
## x22_catch_5yrsImproves slightly         ***
## x22_catch_5yrsStays the same               
## x23_job_secure                          ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.5911 on 4222 degrees of freedom
##   (217 observations deleted due to missingness)
## Multiple R-squared:  0.04974,    Adjusted R-squared:  0.04772 
## F-statistic: 24.56 on 9 and 4222 DF,  p-value: < 2.2e-16

Let me see if there are some that could seem ‘insightful’ for a Rare audience to know. Like knowledge of rules! [Questions 38-40]

Those questions are - Is fishing allowed in the reserve area? (Does fisher X know the rule?) - Would you agree with the statement that most fishers in your community are aware of the boundaries of the fisheries management/managed access area? (Do msot fishers know the rules) - Out of ten fishers in your community, how many would you guess know where the reserve boundary is?

## 
## Call:
## lm(formula = eng_mean ~ x38_reserve_fishing_allowed + x39_ma_boundaries_aware + 
##     x40_reserve_boundaries_aware, data = scaled_sum)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.82012 -0.45885 -0.06591  0.38553  2.01823 
## 
## Coefficients:
##                                           Estimate Std. Error t value Pr(>|t|)
## (Intercept)                              -0.013277   0.028401  -0.467 0.640179
## x38_reserve_fishing_allowed               0.147494   0.021622   6.821 1.04e-11
## x39_ma_boundaries_awareDisagree          -0.271042   0.043462  -6.236 4.95e-10
## x39_ma_boundaries_awareNeither           -0.226942   0.024183  -9.385  < 2e-16
## x39_ma_boundaries_awareNo managed access -0.116381   0.589641  -0.197 0.843543
## x39_ma_boundaries_awareNo reserve        -0.365750   0.178321  -2.051 0.040325
## x39_ma_boundaries_awareStrongly agree    -0.106235   0.027779  -3.824 0.000133
## x39_ma_boundaries_awareStrongly disagree  0.036610   0.065911   0.555 0.578616
## x40_reserve_boundaries_aware              0.012700   0.003262   3.894 0.000100
##                                             
## (Intercept)                                 
## x38_reserve_fishing_allowed              ***
## x39_ma_boundaries_awareDisagree          ***
## x39_ma_boundaries_awareNeither           ***
## x39_ma_boundaries_awareNo managed access    
## x39_ma_boundaries_awareNo reserve        *  
## x39_ma_boundaries_awareStrongly agree    ***
## x39_ma_boundaries_awareStrongly disagree    
## x40_reserve_boundaries_aware             ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.5893 on 3981 degrees of freedom
##   (459 observations deleted due to missingness)
## Multiple R-squared:  0.06071,    Adjusted R-squared:  0.05883 
## F-statistic: 32.17 on 8 and 3981 DF,  p-value: < 2.2e-16

I’ll come back to the regression.

Now, let’s look at a cluster analysis!

Within Engagement and Agreement- We want to see if there are common combinations of answers like “4, 5, 5” for a likert, or something like that.

Now, Clustering

I’m going to explore how many clusters we can see, and try and use a distance-based clustering method first. If possible we want to see common combinations of answers.

The following doesn’t show up in clsuter analysis, but I’d be interested to see… Within 5 questions, if there ae 250 fishers hat say “Yes,” “Slightly agree,” and “No,” we can learn from that.

Basically, it’s going to run 30 different ways of evaluating how many clusters it thinks exist, then tell you the breakdown of what they decide (e.g. “8 algorithms think that there should be 4 clusters”).

NbClust::NbClust()

Now I’ll do K-means clustering to include agreement AND engagement questions.

## K-means clustering with 5 clusters of sizes 701, 842, 936, 1350, 620
## 
## Cluster means:
##   x44_meeting_attendance x45_leadership_position x48_enforcement_participation
## 1              0.4407989              -0.8587732                    -0.3409415
## 2              0.8646081               0.8741093                     0.3539192
## 3              0.9786325              -0.8942308                    -0.8536325
## 4             -0.9733333              -0.8074074                    -0.8807407
## 5             -0.6483871              -0.6629032                    -0.7129032
##   x53_encourage_regulations x61g_fishing_change_behavior x43_ma_benefits
## 1                 1.1298146                     4.523538       0.6918688
## 2                 0.5095012                     4.103325       0.5985748
## 3                -0.7136752                     3.747863       0.5769231
## 4                -0.2911111                     4.050370       0.5296296
## 5                -0.4161290                     3.356452       0.4725806
##   x46_represent_interests x52_ma_benefit_5yrs x61a_current_regulations
## 1               0.7318117           0.6918688                 4.547789
## 2               0.8432304           0.6401425                 3.840855
## 3               0.6880342           0.4476496                 3.951923
## 4               0.6703704           0.4792593                 4.129630
## 5               0.1887097           0.2322581                 2.491935
##   x61f_rights_distribution_fair
## 1                      4.489301
## 2                      3.952494
## 3                      3.717949
## 4                      3.829630
## 5                      2.654839
## 
## Clustering vector:
##    [1] 2 4 1 5 5 4 2 1 2 2 2 1 5 2 5 5 1 2 4 3 5 3 3 2 4 1 1 4 5 3 1 4 2 2 1 4 5
##   [38] 5 5 2 4 4 1 4 2 3 4 4 5 2 1 4 3 3 1 1 2 4 4 3 2 3 2 1 2 1 3 4 1 1 5 3 3 4
##   [75] 5 2 4 4 2 4 2 1 2 4 4 5 5 5 2 4 2 4 2 1 1 4 4 1 4 3 5 3 3 3 3 2 1 1 5 4 3
##  [112] 4 4 3 5 4 4 4 3 4 4 2 3 2 5 3 5 2 4 3 2 1 2 2 1 4 4 3 1 2 4 3 5 4 2 5 3 3
##  [149] 4 4 3 3 3 3 2 1 2 1 4 2 5 2 5 1 4 5 1 3 3 5 4 3 1 4 2 3 4 1 5 3 5 5 1 4 2
##  [186] 2 5 2 1 3 3 3 5 1 5 5 2 5 3 5 2 2 3 4 1 3 5 3 4 1 4 4 4 2 4 4 3 3 3 1 5 4
##  [223] 5 5 4 2 3 5 4 4 2 5 4 5 2 3 2 1 4 3 5 4 3 2 3 4 3 4 4 3 4 4 1 3 4 2 4 4 5
##  [260] 1 1 3 1 4 4 1 1 4 2 5 4 1 3 4 4 3 4 4 4 4 3 2 4 1 4 2 1 1 1 2 4 4 3 4 3 3
##  [297] 1 1 4 3 4 4 4 4 4 4 4 4 3 2 5 4 3 3 3 4 2 1 4 4 1 4 1 5 2 2 2 1 4 2 1 2 3
##  [334] 3 5 2 4 1 4 2 2 2 4 4 4 3 4 5 4 4 1 1 2 4 1 3 1 3 4 5 3 4 4 2 3 2 4 4 1 2
##  [371] 3 3 1 3 1 3 1 4 2 2 3 1 1 4 5 3 5 5 2 3 1 4 3 3 5 4 2 5 5 4 4 2 2 5 1 1 3
##  [408] 3 4 4 4 4 5 2 5 1 2 4 4 4 5 4 2 5 4 1 2 3 1 2 5 1 2 4 3 5 1 1 5 1 4 4 2 3
##  [445] 3 4 3 3 4 4 2 1 4 4 2 3 4 2 4 4 2 1 1 3 4 3 2 2 3 1 1 4 2 1 3 2 3 3 4 5 5
##  [482] 3 1 4 3 3 5 1 2 4 4 2 2 3 2 5 4 3 4 4 4 1 5 1 4 5 2 2 3 3 4 2 5 3 2 3 1 4
##  [519] 3 4 4 4 1 4 4 3 5 4 5 2 5 4 3 4 4 1 4 2 5 4 5 4 5 5 3 5 3 2 4 2 4 4 4 3 3
##  [556] 4 4 3 2 4 4 1 1 2 5 1 2 4 4 2 4 4 1 2 2 4 2 2 2 1 3 1 3 5 2 3 5 5 4 5 4 4
##  [593] 5 2 4 5 4 4 3 5 4 2 5 2 5 3 4 2 4 3 5 4 3 1 2 4 2 3 3 4 4 4 4 2 5 2 1 3 4
##  [630] 4 2 4 1 3 3 4 4 2 2 4 4 2 2 1 5 4 2 2 4 5 4 4 4 3 5 3 2 1 2 4 3 3 1 5 3 5
##  [667] 4 1 1 1 4 4 3 3 2 1 1 3 3 3 1 5 4 4 1 1 4 3 2 5 4 4 3 4 2 5 4 4 2 4 5 4 4
##  [704] 5 5 3 2 5 3 1 5 1 4 4 4 4 2 2 4 3 1 4 4 5 1 3 2 1 3 4 5 3 1 3 4 3 5 3 3 2
##  [741] 3 3 1 5 1 4 4 2 3 3 3 1 3 1 1 3 2 5 4 1 4 3 1 1 3 4 3 5 4 1 4 3 4 2 2 2 4
##  [778] 5 4 1 2 2 1 4 3 5 3 2 3 5 5 3 2 5 3 4 2 5 2 4 3 4 1 4 4 1 3 3 4 5 3 2 1 4
##  [815] 5 3 4 2 2 2 4 3 4 4 1 3 3 1 2 3 1 5 3 3 3 3 4 4 2 2 1 5 3 5 1 2 3 3 5 3 1
##  [852] 4 2 3 1 3 4 5 3 4 2 4 3 4 4 3 3 3 4 2 4 2 1 4 4 5 3 4 4 1 1 5 4 4 4 5 4 5
##  [889] 3 4 2 2 2 3 1 3 3 3 2 4 3 4 5 2 1 2 2 1 4 2 5 3 4 2 3 2 3 4 3 2 4 3 1 5 4
##  [926] 2 4 3 5 1 1 2 3 3 3 4 4 1 2 3 2 4 3 1 5 4 4 4 1 1 5 3 2 2 3 4 2 3 4 3 2 3
##  [963] 3 3 2 5 5 2 4 4 3 3 3 4 3 3 4 1 5 3 1 3 3 3 2 1 1 1 3 3 4 5 3 4 2 3 4 4 4
## [1000] 3 1 2 5 2 3 4 3 1 5 4 3 2 1 3 1 5 4 4 4 1 3 2 3 3 5 3 2 3 3 2 5 2 4 2 1 4
## [1037] 3 4 3 4 1 4 2 4 1 4 3 4 4 3 4 2 2 2 5 2 3 4 3 3 1 4 2 2 3 2 5 4 2 1 3 2 2
## [1074] 4 4 3 2 1 3 1 2 4 4 3 4 1 1 5 2 1 3 1 5 1 3 4 3 3 4 2 3 2 5 1 5 1 3 4 3 1
## [1111] 5 4 1 5 3 1 1 4 5 5 3 4 1 4 2 1 2 5 5 3 2 3 2 5 2 4 4 5 3 3 5 1 5 3 3 1 4
## [1148] 3 4 3 1 2 5 4 5 3 4 4 4 4 1 1 5 1 3 1 4 3 2 1 5 3 4 4 3 2 2 4 3 1 4 4 3 4
## [1185] 4 2 4 5 4 4 2 4 4 1 4 3 5 3 4 3 4 3 2 1 4 5 2 4 2 3 5 2 3 1 2 4 5 4 2 2 3
## [1222] 3 3 4 4 3 4 4 5 2 4 4 5 4 1 2 3 1 2 3 5 5 3 3 3 4 4 2 5 3 4 4 5 4 5 4 4 1
## [1259] 3 4 5 5 2 2 4 4 1 5 4 4 1 1 1 3 3 5 2 2 4 4 3 2 1 5 5 2 3 4 2 4 4 3 5 4 3
## [1296] 3 5 1 3 4 2 4 4 5 3 2 4 4 2 3 4 5 4 5 2 3 4 3 4 4 2 1 3 4 4 5 4 4 2 4 1 1
## [1333] 3 2 2 1 4 1 1 2 4 5 1 4 3 2 2 5 4 4 3 4 4 2 5 2 3 5 3 5 2 3 4 3 2 5 5 4 5
## [1370] 4 2 4 4 1 3 1 3 4 4 4 4 4 1 1 4 2 4 4 5 1 2 5 3 3 3 4 4 2 2 1 4 2 1 1 3 4
## [1407] 3 4 1 1 4 4 2 5 4 2 4 1 5 2 1 2 4 4 2 1 3 4 2 5 4 1 3 3 2 2 3 3 2 4 4 4 1
## [1444] 1 3 2 5 3 4 3 1 3 4 3 5 1 3 2 4 2 1 1 4 3 3 1 5 2 5 3 4 4 1 3 4 4 4 4 1 2
## [1481] 4 4 2 4 3 3 4 5 1 3 2 3 1 2 3 1 2 3 1 1 2 4 1 4 2 5 4 1 5 5 2 2 3 2 1 1 2
## [1518] 5 4 5 4 1 2 5 5 4 4 4 2 3 2 1 5 4 3 4 5 3 1 5 5 4 4 2 5 5 3 4 4 1 1 3 4 3
## [1555] 4 3 3 3 1 1 4 5 3 2 4 1 5 4 3 4 1 2 2 5 4 4 4 3 1 3 4 4 4 1 1 1 3 1 5 4 2
## [1592] 5 4 2 2 4 2 3 2 1 4 5 4 4 2 2 5 4 2 4 2 4 1 1 1 3 3 4 1 2 4 4 4 3 3 2 5 2
## [1629] 3 5 3 1 4 3 1 2 2 5 3 4 4 4 4 2 4 3 4 5 3 4 3 2 1 3 4 1 3 3 4 4 4 4 4 4 4
## [1666] 4 2 4 3 3 1 3 3 4 2 5 4 3 1 2 4 4 4 1 3 3 1 2 4 5 4 2 4 2 4 3 3 4 4 3 1 2
## [1703] 4 1 5 1 3 1 3 2 4 3 4 4 4 5 5 3 4 5 3 3 2 2 4 1 2 3 2 4 1 4 4 5 5 3 1 5 4
## [1740] 5 3 3 4 4 4 3 3 3 4 4 4 4 5 3 2 4 5 1 2 5 2 3 4 4 4 5 4 4 4 3 5 1 3 5 5 4
## [1777] 3 4 4 4 1 2 4 5 4 1 4 2 3 3 4 4 5 3 4 3 1 4 4 3 1 2 2 2 4 2 2 1 3 1 4 2 4
## [1814] 4 3 5 5 4 4 4 4 1 5 4 1 1 4 2 4 1 5 3 2 4 5 5 5 3 5 1 1 2 1 3 5 2 5 5 2 5
## [1851] 4 1 1 1 4 2 3 3 4 4 2 4 2 4 4 1 3 5 3 3 4 3 4 4 4 2 3 3 4 2 1 2 4 3 4 1 1
## [1888] 3 2 1 4 3 4 1 4 3 2 1 4 4 3 2 3 2 5 4 4 1 1 3 2 4 3 1 1 1 4 5 1 3 1 4 3 1
## [1925] 4 5 2 3 4 2 1 4 5 2 1 4 4 4 5 5 3 3 2 4 1 5 4 1 2 1 3 4 4 3 1 3 4 4 3 3 3
## [1962] 2 4 2 2 3 3 5 1 5 3 4 1 3 4 1 1 2 5 1 1 3 1 1 4 3 4 2 3 5 2 2 2 3 4 2 4 4
## [1999] 1 3 1 2 4 3 3 3 1 4 3 4 4 4 2 2 3 2 1 4 1 5 4 2 3 3 1 4 4 4 5 2 5 4 5 3 4
## [2036] 1 2 4 4 5 1 4 4 2 4 5 4 4 4 1 3 4 5 5 2 2 4 3 3 4 4 4 3 5 5 1 4 2 4 5 1 2
## [2073] 3 3 4 3 5 3 1 4 4 2 5 3 4 4 1 4 4 2 2 4 4 1 4 5 4 2 1 4 3 4 4 3 5 1 3 2 1
## [2110] 4 4 5 3 4 1 3 5 4 1 2 2 4 4 4 5 3 4 2 4 5 5 4 4 1 4 4 3 4 2 1 2 4 4 5 4 4
## [2147] 4 4 4 2 4 1 4 1 5 4 1 5 5 3 4 2 3 1 2 4 1 4 5 1 4 4 1 1 3 1 2 1 3 3 5 4 2
## [2184] 5 4 4 2 3 4 2 3 4 2 2 5 4 4 5 4 4 4 3 1 5 4 2 3 2 2 2 4 4 3 2 3 3 2 5 4 5
## [2221] 2 4 3 3 2 5 4 2 3 1 4 4 1 4 2 4 2 3 2 3 5 4 4 2 4 5 4 3 2 2 4 5 3 2 3 1 3
## [2258] 2 2 3 5 4 3 5 3 4 3 1 4 2 1 4 2 2 3 1 5 3 1 4 1 2 3 4 5 4 4 1 3 3 2 4 2 3
## [2295] 5 4 3 3 4 5 4 4 5 3 3 5 4 2 2 4 4 3 2 1 3 1 1 4 1 4 3 2 2 3 2 5 4 2 4 5 3
## [2332] 4 2 1 1 2 5 2 4 1 4 3 5 2 4 1 4 1 2 2 2 4 4 1 3 3 4 1 2 4 3 5 4 3 5 5 4 4
## [2369] 3 4 2 5 1 2 4 2 3 3 4 5 4 4 4 2 5 1 5 3 1 2 2 2 4 2 1 3 4 3 3 5 4 4 3 4 3
## [2406] 2 4 3 2 2 4 4 3 2 5 2 2 1 2 3 5 3 2 2 1 1 2 2 1 2 1 1 3 5 1 5 5 5 2 1 1 4
## [2443] 3 3 3 2 4 2 4 4 3 4 3 2 4 3 4 1 5 1 3 1 3 5 3 4 1 3 2 4 3 2 4 1 2 2 1 4 4
## [2480] 5 4 3 3 5 1 2 2 3 3 1 2 1 3 3 3 2 1 2 3 1 1 5 4 2 2 3 1 1 4 4 3 5 5 4 1 2
## [2517] 2 4 5 2 1 4 1 4 1 2 3 5 3 1 2 4 4 4 2 2 4 4 5 5 3 5 3 2 4 4 4 3 4 4 3 4 4
## [2554] 5 3 3 3 3 4 5 1 4 4 4 3 1 1 4 3 3 2 1 3 5 3 3 4 3 4 3 1 5 3 3 4 3 4 5 1 4
## [2591] 2 1 3 3 3 3 5 2 2 4 4 4 1 4 4 2 3 4 4 4 4 3 1 3 5 1 2 4 2 4 4 5 3 3 4 4 3
## [2628] 3 2 5 1 1 2 4 3 4 4 3 2 3 5 3 5 2 2 1 4 2 2 5 3 4 2 4 3 3 4 4 3 1 1 3 5 3
## [2665] 3 4 4 2 1 1 1 4 2 1 2 4 4 1 3 1 1 5 5 4 5 2 3 4 4 3 2 4 4 5 4 4 2 5 1 3 2
## [2702] 1 4 3 3 5 2 1 4 4 4 2 4 3 3 5 5 1 5 1 1 5 4 2 4 2 1 5 2 2 1 2 5 1 4 3 3 5
## [2739] 2 4 4 1 3 2 3 2 5 3 2 5 2 3 4 4 1 1 3 4 5 4 4 4 5 5 4 4 1 5 3 2 2 4 4 3 3
## [2776] 5 4 5 2 3 3 1 4 2 3 4 5 4 2 4 4 2 2 4 5 5 1 4 3 2 3 1 4 2 4 4 3 4 3 1 2 3
## [2813] 4 1 5 3 5 2 3 2 5 3 4 4 3 4 4 2 3 2 4 1 3 4 2 1 4 3 2 2 4 5 2 2 4 3 3 2 4
## [2850] 2 3 3 2 4 2 4 5 2 3 2 4 5 5 5 3 3 1 4 4 5 3 3 5 1 4 4 2 4 2 4 3 2 4 3 2 4
## [2887] 1 4 4 5 2 5 4 4 4 4 3 3 2 4 4 2 1 1 5 3 3 2 5 5 5 3 3 1 5 3 5 5 4 4 5 5 4
## [2924] 2 3 2 1 4 2 2 3 4 3 3 1 5 4 4 4 3 3 4 5 3 1 5 4 1 3 4 4 2 3 2 2 1 4 3 2 5
## [2961] 5 1 4 1 5 3 1 1 1 1 1 5 4 2 1 5 3 3 4 2 4 1 5 3 1 2 5 2 4 3 1 4 4 4 5 1 1
## [2998] 1 5 4 3 1 5 5 3 4 1 2 5 3 3 2 5 4 3 5 1 1 1 4 2 2 4 2 2 1 4 3 4 4 3 5 1 2
## [3035] 4 5 2 2 5 5 3 4 5 3 4 4 2 2 4 3 2 4 1 1 5 1 4 2 3 3 4 1 4 5 1 4 4 3 4 5 1
## [3072] 1 1 1 2 2 4 5 4 3 4 2 5 4 4 1 2 3 2 1 1 4 4 2 2 4 4 1 4 3 1 4 3 1 5 4 4 4
## [3109] 4 3 4 4 2 4 4 3 3 5 1 3 5 1 3 1 1 3 3 4 1 3 2 2 4 1 2 5 3 3 4 2 3 5 1 1 3
## [3146] 5 5 3 4 2 5 4 4 2 1 1 1 4 4 4 2 3 2 1 5 2 2 4 4 2 1 2 3 1 4 5 2 3 2 2 3 3
## [3183] 4 3 3 2 4 5 1 2 5 5 2 2 1 5 4 5 3 4 2 3 1 4 4 4 4 2 5 3 4 2 2 2 3 5 4 2 2
## [3220] 4 3 3 4 4 3 5 5 4 3 5 3 4 5 3 4 5 4 1 5 4 1 4 3 3 3 2 1 2 4 4 5 3 4 2 1 4
## [3257] 1 2 1 2 3 4 4 4 4 4 5 1 2 1 2 3 2 5 2 3 3 1 4 4 2 2 5 4 4 3 2 1 2 1 3 3 4
## [3294] 1 4 3 4 3 3 3 3 3 3 3 4 4 4 1 2 5 4 1 1 2 4 4 2 2 4 4 3 4 4 2 5 4 3 1 3 2
## [3331] 4 3 2 4 1 4 2 5 1 4 2 1 5 4 3 2 5 4 1 4 2 4 2 5 2 2 2 4 4 5 3 2 1 1 4 4 4
## [3368] 5 4 3 5 3 3 4 4 4 1 2 1 2 2 2 1 1 4 4 2 4 3 4 4 5 5 1 1 1 3 1 4 5 4 5 2 4
## [3405] 2 2 1 4 5 1 1 1 4 2 4 1 2 4 2 2 3 5 1 5 4 2 4 1 5 3 5 2 2 4 5 5 2 4 3 1 5
## [3442] 2 4 2 3 3 1 2 3 4 3 2 2 1 3 5 4 4 4 3 5 4 5 1 1 4 1 3 3 1 4 4 4 4 3 3 2 4
## [3479] 3 4 2 5 1 1 5 4 2 4 3 3 2 4 4 4 2 2 2 2 4 4 4 1 5 1 2 2 5 4 4 4 5 2 5 3 2
## [3516] 1 2 1 4 3 5 5 3 2 1 1 4 3 4 3 2 3 5 4 1 5 2 4 3 3 4 5 5 4 3 4 1 4 5 2 2 3
## [3553] 4 4 4 1 4 2 2 2 4 1 3 4 2 2 3 2 3 5 4 2 1 4 2 4 3 2 1 2 3 4 5 2 3 3 4 1 5
## [3590] 1 1 4 2 3 4 2 2 1 4 3 4 2 1 4 4 4 4 4 4 4 4 5 1 5 5 4 4 3 1 4 1 3 3 3 4 4
## [3627] 2 4 4 3 5 4 4 5 4 2 3 4 1 2 2 5 2 3 1 4 2 2 3 4 4 4 3 5 1 2 4 4 2 4 4 4 2
## [3664] 4 2 4 2 2 4 3 5 4 5 3 1 5 2 2 1 5 5 1 2 1 5 2 4 3 1 4 4 5 3 4 2 4 4 4 4 3
## [3701] 1 5 3 4 1 4 5 1 2 3 4 4 4 3 4 2 3 4 2 2 5 4 2 4 2 5 3 1 3 3 2 3 5 1 4 1 1
## [3738] 3 2 3 4 3 2 1 3 2 2 3 2 4 1 4 1 2 3 5 5 4 4 5 4 1 4 2 4 3 4 3 2 4 2 2 1 5
## [3775] 2 2 3 2 3 2 2 3 3 4 3 2 4 2 5 2 1 2 3 1 3 4 4 4 4 5 5 2 4 2 1 3 4 2 2 1 4
## [3812] 1 2 5 2 2 4 5 3 5 1 3 4 2 1 1 3 4 4 1 5 4 1 1 1 3 2 1 3 2 4 4 2 3 5 4 4 4
## [3849] 1 3 4 3 2 1 2 2 2 5 2 2 4 1 2 2 1 4 4 5 1 4 2 3 4 3 2 4 3 4 5 1 3 5 4 5 4
## [3886] 4 2 2 4 4 2 1 2 5 5 4 2 1 4 4 4 5 4 4 4 1 5 1 1 4 5 1 1 2 4 4 1 4 4 4 4 4
## [3923] 5 5 4 4 2 1 4 3 1 2 5 5 4 5 5 2 4 5 2 3 4 1 4 4 2 1 4 1 1 4 2 2 1 5 4 3 1
## [3960] 4 5 4 2 3 2 3 3 1 1 2 3 1 5 3 3 4 5 2 1 5 1 5 4 4 3 3 1 4 4 4 5 5 2 3 4 3
## [3997] 4 4 2 5 4 4 4 4 4 2 3 4 2 5 1 4 2 3 4 5 3 4 4 4 2 4 3 2 2 5 1 5 4 2 4 2 4
## [4034] 1 2 4 4 3 2 3 2 4 3 5 5 2 5 2 3 2 2 3 4 4 4 4 4 2 1 4 3 3 5 5 5 4 3 3 5 4
## [4071] 5 2 4 4 2 3 4 2 3 2 4 4 2 1 5 1 4 1 2 4 2 3 2 4 4 4 4 3 1 4 3 5 4 5 3 4 4
## [4108] 4 3 3 4 3 4 2 5 2 5 5 1 5 3 3 3 4 3 3 3 2 1 3 4 4 4 2 4 1 3 3 1 5 3 4 4 1
## [4145] 3 4 4 2 4 5 4 1 5 3 3 4 1 3 5 4 5 3 4 2 5 4 3 4 1 3 3 3 3 4 2 4 1 2 5 4 1
## [4182] 5 3 4 1 4 5 2 4 4 4 4 3 5 2 2 3 2 2 2 2 3 5 5 2 4 4 4 5 4 4 3 3 2 2 3 1 3
## [4219] 5 3 3 4 2 4 1 4 5 4 4 5 1 5 3 4 1 3 1 5 4 3 4 4 3 3 4 3 4 5 3 3 4 5 3 4 3
## [4256] 2 3 4 4 2 2 5 5 2 4 1 1 5 5 3 1 1 4 3 5 5 4 3 3 5 5 1 4 2 1 1 4 3 1 5 3 3
## [4293] 2 4 1 3 5 4 5 3 2 5 1 1 3 5 2 4 3 2 1 3 4 4 4 4 2 3 4 1 1 1 4 4 3 1 1 4 4
## [4330] 4 4 3 4 4 3 2 4 1 3 2 1 3 4 2 2 4 4 3 5 5 4 4 4 1 3 4 4 5 1 4 2 3 4 5 5 3
## [4367] 3 3 4 1 2 1 5 3 2 3 1 1 1 4 2 2 2 4 5 1 1 1 4 3 4 4 1 4 4 3 4 1 2 1 3 3 2
## [4404] 1 2 4 2 3 3 2 2 2 2 2 3 5 5 3 1 3 5 4 2 5 4 1 1 4 2 5 2 1 3 4 4 5 3 1 3 2
## [4441] 3 3 3 2 4 3 3 1 4
## 
## Within cluster sum of squares by cluster:
## [1] 2910.662 3338.550 3249.987 4476.612 3886.760
##  (between_SS / total_SS =  39.2 %)
## 
## Available components:
## 
## [1] "cluster"      "centers"      "totss"        "withinss"     "tot.withinss"
## [6] "betweenss"    "size"         "iter"         "ifault"

##   agreement_engagement_km.size x44_meeting_attendance x45_leadership_position
## 1                          701              0.4407989              -0.8587732
## 2                          842              0.8646081               0.8741093
## 3                          936              0.9786325              -0.8942308
## 4                         1350             -0.9733333              -0.8074074
## 5                          620             -0.6483871              -0.6629032
##   x48_enforcement_participation x53_encourage_regulations
## 1                    -0.3409415                 1.1298146
## 2                     0.3539192                 0.5095012
## 3                    -0.8536325                -0.7136752
## 4                    -0.8807407                -0.2911111
## 5                    -0.7129032                -0.4161290
##   x61g_fishing_change_behavior x43_ma_benefits x46_represent_interests
## 1                     4.523538       0.6918688               0.7318117
## 2                     4.103325       0.5985748               0.8432304
## 3                     3.747863       0.5769231               0.6880342
## 4                     4.050370       0.5296296               0.6703704
## 5                     3.356452       0.4725806               0.1887097
##   x52_ma_benefit_5yrs x61a_current_regulations x61f_rights_distribution_fair
## 1           0.6918688                 4.547789                      4.489301
## 2           0.6401425                 3.840855                      3.952494
## 3           0.4476496                 3.951923                      3.717949
## 4           0.4792593                 4.129630                      3.829630
## 5           0.2322581                 2.491935                      2.654839

Summary of Clusters

x44_meeting_attendance	x45_leadership_position	x48_enforcement_participation	x53_encourage_regulations	x61g_fishing_change_behavior	x43_ma_benefits	x46_represent_interests	x52_ma_benefit_5yrs	x61a_current_regulations	x61f_rights_distribution_fair
0.4407989	-0.8587732	-0.3409415	1.1298146	4.523538	0.6918688	0.7318117	0.6918688	4.547789	4.489301
0.8646081	0.8741093	0.3539192	0.5095012	4.103325	0.5985748	0.8432304	0.6401425	3.840855	3.952494
0.9786325	-0.8942308	-0.8536325	-0.7136752	3.747863	0.5769231	0.6880342	0.4476496	3.951923	3.717949
-0.9733333	-0.8074074	-0.8807407	-0.2911111	4.050370	0.5296296	0.6703704	0.4792593	4.129630	3.829630
-0.6483871	-0.6629032	-0.7129032	-0.4161290	3.356452	0.4725806	0.1887097	0.2322581	2.491935	2.654839

There are 5 clusters!

Sizes: 701 842 936 1350 620

Comparing clusters

1 and 2 seem similar, with high levels of A and E - 44: 0.44 (C1) and 0.86 (C2) Cluster 2 attends more meetings - 45 is a big swing, cluster 2 is 0.87, and C1 is -0.85. These are the leaders! - 48: -0.44 and 0.35 is a decent difference in enforcement BUT C1 is more likely to encourage regulations (1.21 v 0.5) -61 is closer (4.5 and 4.1)

Within agreement they’re more similar. but C1 is higher average in each of the 5 questions. Less leaders, but agree more across the board.

3 and 4 - 44: C3 has a 0.97, and C4 has a -0.97 45 is similar (within .01) Same for 48 53 (encourage regulations) is -0.7 (C3) and -0.29 (C4) 61g change behavior C3 is the 2nd lowest at 3.7, and c4 is in third at 4.05 Within agreement, they’re averages are similar across all questions, with C3 a little bit lower for most, except ma_benefits

More than anything, C3 ATTENDS a whole lot more than C4. but C3 is the least likely of any group to encourage regulations. Perhaps we’re seeing creative fishers here

C5 is looking at 2/5 for questions 61a and 61f, and is also by far the lowest for 46 and 52. within agreement, only in MA benefits are they somewhat close to other clusters. They do have a few more leaders than other clusters though. But overall, very much not attending

## Warning: Removed 10 rows containing missing values (geom_point).

Clusters with question 64: How wrong is fishing in the reserve?

## Warning: Removed 1336 rows containing missing values (geom_point).

Table version:

Morality

cluster_no	x64_wrong_fishing_reserve	n
1	5	272
1	4	197
1	3	65
1	2	6
1	1	3
1	NA	158

So, kind of everyone thinks its pretty wrong, huh. Also a ton of NAs, hard to limit that noise

## Warning: Removed 3 rows containing missing values (geom_point).

Right away this is really interesting!

Cluster 1 Only 1 observation is below -1 in agreement. Most of the fully-agreeing fishers are in cluster two - but not all! It excludes the agreeing, but rather unengaged fishers. Similarly, it omits the super-engaged. We can describe C1 (so far) as fishers who are very on board with rules, but are a bit above average engagement. Centers = (0.448207772, 0.3930240) (Agreement, Engagement) N = 701

Cluster 2 These are all the super-engaged fishers. Agreement ranges from 1 (the maximum) to -1.4. Center = (0.156193496, 0.8606600) (Agreement, Engagement) N = 842

Cluster 3 This is the largest spread of any cluster. We have some fully-agreeing fishers, and at least 1 or 2 with agreement below 1.5. Engagement grows up to 1.5, which is mostly reserved for the high-engaging clusters. With means SO close to 0, we see this as a the most average cluster. A bit less engaged, and agreeing with as much as the average fisher from the study. Center = (0.002805289, -0.1748106) (Agreement, Engagement) N = 936

Cluster 4 Its center is very close to Cluster 3, but we can see clear differences when we look at the spread. These folks do NOT engage. While there as unengaged as the Resistant fishers, but most actually agree quite a lot. This includes some with A = 1, the max. Center = (0.052635489, -0.4098127) (Agreement, Engagement) N = 1350, largest cluster

Cluster 5 - Resistant This is the classically known group of unengaged, disagreeing fishers. Many of the outliers are from Honduras, but the country variable doesn’t explain the entire cluster. They all disagree, but to varying levels. We’ll look at their distinguishing characteristics. Center = (-0.837729407, -0.4569608) (Agreement, Engagement) N = 620, smallest cluster.

Notes: - Resistant fishers are far more present in BRA, HND and PHL. However, HND and BRA both spread from the lowest to highest engagement possibilities. This mean there is OPPORTUNITY to be a leader, enforce, etc.

Plotting clusters by gender Female mean = (-0.06947793, -0.03637067)

Male mean = (0.04032597, 0.02179388)

This was not the case before! Were all within 0.1 of zero, to be clear. But when we add Brazil, the mean for men increases

• Agreement-onyl clusters, not in use.* Comparing the clusters, by raw questions:

7 machines said there are two clusters, I’m going to run this as well and compare a little:

To do so, I’ll make 4 df, one for each cluster.

## [1] 0.393024

## [1] 0.4482078

## [1] 0.3411713

## [1] 0.5396342

## [1] 0.86066

## [1] 0.1561935

## [1] 0.9495807

## [1] 0.1382421

## [1] -0.1748106

## [1] 0.002805289

## [1] -0.2266834

## [1] 0.04425969

## [1] -0.4098127

## [1] 0.05263549

## [1] -0.4500038

## [1] 0.1214331

We see some slight changes between means and medians, BUT we care about the outliars, (mostly the very low engagement and agreement.) So, we’ll use the mean

Cluster-specific maps

Cluster 1 Notes:

A subset of quite agreeing, and somewhat engaged fishers.

Cluster 2 Counts: BRA 25, HND 29, IDN 33, MOZ 322, PHL 422

Notes: Committed fishers, easily the highest engaging group. Like C1, has a good chunk of high-agreeing fishers also. Those top engagement scores seem to come from nearly all countries (less from Honduras.)

Cluster 3 Notes:

Just as with the full graph, this has the most spread, centered very close to (0,0.) on Engagement we see -1 up to 0.5 and -0.5 to 1 with agreement.

BRA 67, FSM 1, HND 34, IDN 18, MOZ 25, PHL 783, PLW 8

Cluster 4

## Warning: Removed 1 rows containing missing values (geom_point).

otes: We see that this is determined by low engagement, but decent to high agreement. The cluster ranges from -1 to 1 on agreement. These folks are OK with rules for the most part. However, they don’t take part in the management process. More than ⅓ of all Brazil’s respondents are in this cluster.

BRA 283, FSM 6, HND 126, IDN 32, MOZ 167, PHL 731, PLW 5

Cluster 5

## Warning: Removed 9 rows containing missing values (geom_point).

Resistant Fishers

With a mean at (-0.5, -.8) this is a disagreeing, unengaged group, including some crazy disagreeing outliers, from Honduras. There are a chunk of observations above 0 in engagement, but only 1 on the agreement axis. N = 620, smallest cluster.

Country-specific cluster maps

Philippines: We see this breakdown:

1. 494 (19%)
2. 422 (16%)
3. 783 (30%)
4. 731 (29%)
5. 134 (5%)

Indonesia:

## Warning: Removed 1 rows containing missing values (geom_point).

Counts C1: 10 (7%) C2: 33 (23%) C3: 18 (12%) C4: 32 (22%) C5: 53 (36%) Honduras

## Warning: Removed 1 rows containing missing values (geom_point).

Counts:

C1: 6 (21%) C3: 1 (4%) C4: 6 (21%) C5: 15 (54%)

Mozambique

## Warning: Removed 7 rows containing missing values (geom_point).

Exploring relations between clusters and other variables (Plus hypotheses)

Engagement-Only Clustering

This is not part of our continuing analysis

So R has suggested that there are 3 meaningful clusters for engagement.

For now, I run the clusters, and we can go back.

Looking at 3 clusters:

OK! That’s a whole lot of information.

I’ll look at a couple key factors

• size
• cluster (which cluster each observation is in)

Size of clusters: [1] 1277 878 959

Now that these clusters are attached to each observation, let’s visualize it

This is great, let’s map them over country and reserve as well.

Correlation plots with corrplot()

When running a correlation of all the Engagement questions, we see very little correlation across the different questions. We see correlations around 0 and 0.3, with no negative correlations. We find the largest correlation between questions 45 and 48 (0.5).

Let’s do the same with agreement, and all compliance questions together.

Here we see really similar results to the Engagement corrplot. There are no dark blue or large circles. Again, we see no negative correlation, a lot of results between 0 and 0.2. 61(f) (fairness of distribution) and 61(a) (benefit of current regulations) have a slightly lager correlation, at 0.5

Let’s look at agreement AND engagement. This will look messy at first.

Here, we see nothing new, which is interesting! There are no strong correlations across Engagement and Agreement. This suggests that they are genuinely different measurements of compliance. Also of note, there are ZERO negative correlations between an agreement question and an engagement question. I thought that we might see something below zero, but all of these questions track towards compliance. Quite useful stuff.

Bar Graphs for each A & E Question

Engagement

1. Question 44

Question 44: Mean = 0.138, but nearly all respondents said Yes or No. Around 1200 responses did not attend meetings, will around 1600 do attend.

Question 45

Question 45: Mean = -0.551. This mean shows that 3/4 of fishers responded “No” Few fisher households included someone with a leadership position within the fishery.

Question 48

Question 48: Mean = -0.492. Nearly 3 out of 4 fishers do not participate in enforcement.

A quick pause from graphs!

Let’s see the combination counts for question 45 and 48

• Leaders who don’t enforce: 192
• Enforcement volunteers without leadership roles: 250
• Households who do both: 283 -Households who no neither: 1770

Question 53: Mean = -0.071. This is just below the “neutral” survey option. We see a very similar totals in the “strongly agree” and “strongly disagree.” Our mean skews negative because around 100 more fishers wrote “disagree” relative to “agree.”

Question 61(g) mean = 3.955. This is the first mean that really skews high. With a mode of 4, most respondents were at least “willing” to change their fishing behavior for the reserve structure.

Here’s a check for normal distrbution: It looks pretty normal, even with the high amount of “4” responses

Moving onto Agreement

43: Mean = 0.556. Nearly no one answered with the negative response.

*Question 46: Mean = 0.667. Huge skew towards the positive response.

We saw that variables 46 and 47 are highly correlated, and have similar wording. We find them to NOT be independent from one another, and exclude 47 from the index of Agreement.

*Question 52: Mean = 0.506. For the first time, we see a lot of “unsure” answers. This likely has to do with uncertainty in projecting 5 years in the future.

Question 61(a): Mean = 3.93, clear positive skew, and the mean seems to represent the mode and median

61(f): Mean = 3.805. More neutral responses that strongly agree, but very few disagreeing fishers.

Question 62: Mean = 0.525

We do NOT include this question in the index, but will use it to create a graph of “admitting noncompliance”

Question 64: Mean = 3.957. Another mean between 3 and 4, with very small numbers of below-neutral responses

Visualizing Variables outside of compliance, against clusters:

Starting with Enforcement

“Who is responsible for enforcement in your community? 1. Fisheries Management Body 2. National Government 3. Subnational Government 4. Other (please state) 5. There is no enforcement system”

What do we see??

Dyl hypothesis: National government would show us less engaged fishers

• Cluster 1’s largest porpotion is within the smaller “Myself” group
• Cluster 1 is smaller for FMP than sub-national and national government (suprised me)
• “No enforcement” gives us the largest contingent of resistant fishers
• Cluster 3 (who attend but do not encourage rules) is tiny within “No Enforcement”
• Cluster 3 is large when the subnational government is responsible for enforcing (perhaps they attend, but don’t see any impetus on fishers to do anything more)
• C2 (leaders, very engaged) are are the largest contingent when FMB is responsible for enforcement (this makes sense)

Let’s get some count data (classic post-visualization move)

Enforcement

cluster_no	Fisheries Management Body	Myself	National Government	No enforcement system	Other	Subnational Government
1	251	45	199	14	6	166
2	603	16	123	4	16	60
3	481	14	125	14	8	270
4	511	28	248	67	12	445
5	248	14	123	89	14	88
Total	2094	117	818	188	56	1029

Means of Agreement and engagement, based off enforcement type

Comparing “Subnational government” and “No enforcement system”

## Warning: Removed 10 rows containing missing values (geom_point).

Summary Statistics

Variable	N	Mean	Std. Dev.	Min	Pctl. 25	Pctl. 75	Max
eng_mean	188	-0.485	0.5	-1.559	-0.829	-0.207	1.586
agree_mean	188	-0.498	0.73	-2.207	-0.983	0.101	0.98
cluster_no	188
… 1	14	7.4%
… 2	4	2.1%
… 3	14	7.4%
… 4	67	35.6%
… 5	89	47.3%
x49_enforcement_responsible	188
… No enforcement system	188	100%
x51b_fishers_reserves	184	2.81	3.613	0	0	5	10
country	188
… BRA	142	75.5%
… HND	43	22.9%
… MOZ	3	1.6%

No enforcement system lacks engaged folks, perhaps because they cannot volunteer for enforcement There is more high agreement among fishers who believe that their National Government is responsible for fishery enforcement The very-unengaged (below -1) have an even spread across the 3 types. Among the No enforcement: Variable N Mean Std. Dev. Min Pctl. 25 Pctl. 75 Max eng_mean 188 -0.485 0.5 -1.559 -0.829 -0.207 1.586 agree_mean188 -0.498 0.73 - 2.207 -0.983 0.101 0.98 47% in C5, and 35% in C4. 75% from Brazil, 22% Honduras. There are a few outliars who engage heavily, and at least 1 who completely agrees. However, these folks are mostly resistant. 1 SD above the mean in right around (0,0)

Color-blind Dyl here, uh oh! I’m going to need to see some tables FMP has fewer fishers who heavily disagree. That said, I can’t see much from these graphs so far, other than the interpretations from the bar graph.

OK! Dad wants to see boat ownership Hypothesis: Fishers without boats are less likely to engage

Visual Interpretation

• When the collective owns a boat, C4 dominates, C2 (super-engaged) shrinks
• When fishers share boats “as an employee” (meaning their boss owns a boat?) there are more resistant fishers.
• Fishers who own their own boat have a larger contingent of Cluster 1
• Renters are the most “super-engaged” (C2) and least resistant (C5)

Let’s break this down by country, to check if boat status and country are heavily correlated

A lot of renting on Honduras! Mozambique and Brazil have a lot of collective ownership Brazil also has a lot of boat owners Mozambique has the highest percentage of fishers who fish without boats, even more than Brazil (where I know many fishers collect crabs and other invertebrates)

Hypothesis: Boat employees fish industrially

OK, too hard to know. 231 NA, 15 not NA. No correlation

Visualizing Need + Food Security

HHS questions that relate to Need

• Question #23: “Do you believe that the job of a fisher is secure in the future?”
  
• Question #26: “Does your income provide enough to save?”
• Question #28: “Does your household take out loans from fish buyers or traders?”
• Question #29: “To cover family needs your household income is…”
• Question #54: “How do you rate the last year in terms of food availability?”
• Question #55: “Consider the following statement:”I worry about not having enough food for everyone in the household” Was that often, sometimes or never true for you in the last 12 months?"
• Question #59: “Are you confident that you will be able to procure enough food for you and your family for the next 12 months?”
• Question #60: In the last 12 months, how often did your household eat fish?

Starting with 55 and 56

This is going to be a bunch of visuals, and some comparison to see what’s actually shown in the data.

Dyl’s hypotheses: - Fishers with limited funds are less engaged, but vary evenly across agreement - Fishers who eat lots of seafood are more engaged than those who don’t - Cluster 1 (agreers) make more money than Cluster 2 (super-engaged)

## Warning: Removed 10 rows containing missing values (geom_point).

Cluster 1: 66% job secure Cluster 2: 72% job secure Cluster 3: 55% job secure Cluster 5: 46% job secure

Question 29 “To cover family needs your household income is…” [Sufficient, Tight, Insufficient]

## Warning: Removed 10 rows containing missing values (geom_point).

Let’s look at some count numbers (percentages are numbers!)

C1: 27% Insufficient, 15% sufficient, 57% Tight 671 total C2: 25% Insufficient, 21% Sufficient, 54% Tight 780 total C3: 26% Insufficient, 10% Sufficient, 63% Tight 887 total C4: 11% Insufficient, 17% Sufficient, 70% Tight 1241 total C5: 21% Insufficient, 21% Sufficient, 58% tight 584 total

Since need (so far) isn’t correlating much with any particular level of compliance, let’s see how people who are dependent on seafood comply:

## Warning: Removed 10 rows containing missing values (geom_point).

Even looking at counts, we’re seeing another variable without any correlation - this si kinda crazy!

What if - nothing really drives compliance behaviors other than simply attitudes about compliance?

… There are plenty more variables to explore before we truly wonder that, though.

Let’s look at x55_worry_food

## Warning: Removed 10 rows containing missing values (geom_point).

Alright, now we see some spread.

Those who “often” worry about food are MORE engaged and also the most resistant.

This is challenging because there is a far larger N within “sometimes,” but we can compare the differences of “Never” and “Often.” Never has some very disagreeing fishers, who engage a little (willing to change behavior, likely)

Those who often worry about food have a lot of somewhat engaged, rather agreeing folks. Let’s do some stats here. He

Corrplots for many variables

## Warning: Unreplaced values treated as NA as .x is not compatible. Please specify
## replacements exhaustively or supply .default

## Warning in x50_ma_punishment != c("No management", "NA", "na"): longer object
## length is not a multiple of shorter object length

## Warning: Unreplaced values treated as NA as .x is not compatible. Please specify
## replacements exhaustively or supply .default

## Warning in cor(first_correlate_df, use = "pairwise.complete.obs", method =
## "spearman"): the standard deviation is zero

## Warning in cor(first_correlate_df, use = "pairwise.complete.obs", method =
## "spearman"): the standard deviation is zero

## Warning in cor(first_correlate_df, use = "pairwise.complete.obs", method =
## "spearman"): the standard deviation is zero

## Warning in cor(first_correlate_df, use = "pairwise.complete.obs"): the standard
## deviation is zero

Correlation with Engagement/Agreement:(N = Negative, P = Positive, options of strength = c(week, moderate, strong)) [E/A] - x11e (P, weak) [E]

• x61c (P, weak) [E]
• x61d (P, weak) [E]
• x61h (N, weak) [E] Weak, negative correlation with Engagement: x55, x11b, x11d, x11f, x11h, x11i, x39, x50, x61h, x63, x64 Agreement:
• x61c (P, weak) [A]
• x61d (P, weak) [A]
• x61h (P, weak) [A]
• x61i (P, weak) [A]
• x64 (N, moderate) [A]
• x11h (N, moderate) [A] x39 (N, moderate) [A] Weak negative correlation with agreement: 11b, 11c, 11d, 11f, 11g, x62, x63

Other interesting correlations - x64 (wrong to fish in the reserve) and x11h (income from extraction) (P, moderate) - x61d (strong enforcement) and x11h (income form extraction) (N, moderate) - Strong negative correlation between 11i and 11h (tourism and extraction incomes) - x64 has negative correlations with x61c and x61c (community participation and strong enforcement) - x61c through x61 i have moderate positive correlations

Making a table of all these variables, by cluster

Let’s compare question 19 across the axes and the clusters

## Warning: Removed 10 rows containing missing values (geom_point).

## Warning: Removed 10 rows containing missing values (geom_point).

Eye-test notes: - Declined a lot has most of the resistant outliars, has less average fishers AND less of the high-agreement, low-engagement folks - Improved Slightly: not too many in full-agreement with rules, but very even across the engagement axis. Also miss thing high-agreement, low-engagement fishers. - Stayed the same has a limited amount of high-engaging fishers, we see a lot of high-agreement.

For me, I’d like to see some counts

Here’s a breakdown of the ecology question:

• Declined a lot: 871

• Declined slightly: 1855

• Improved heavily: 27

• Improved slightly: 372

• Stayed the same: 1169

Cluster 1 saw mostly declines (466/677) Cluster 2 has the most who saw slight improvements Cluster 3

I just want to check the counts of x19 by country